Marc Marschark

A Reexamination of the Role of Imagery in Learning and Memory

Five experiments were conducted to examine whether the superior recall of concrete over abstract words might be better accounted for in terms of relative differences in the processing of relational and distinctive information rather than redundant verbal and imaginal memory codes. Concrete and abstract word pairs were presented in the standard paired-associated learning task or under conditions intended to affect the nature and extent of relational processing between pair members. Concreteness effects were attenuated or eliminated when relational processing was prevented at encoding (Experiments 3, 4, and 5) or when the use of encoded relations within pairs was prevented at recall (Experiments 1, 2, and 3). The results indicated the viability of an account of concreteness effects in paired-associate learning based on the joint functions of distinctive and relational information. They also remove theoretical constraints imposed on imagery theories by the incorrect assumption of a uniform presence of concreteness effects in memory for word lists.

Imagery and organization in the recall of prose

Abstract A theoretical puzzle in research on imagery and memory is that even when concrete and abstract sentences can be shown to receive equal amounts of integrative processing, concrete ones still are remembered better than abstract ones. Three experiments examined recall of concrete and abstract sentences that composed paragraphs comparable in comprehensibility, syntactic structure, and conceptual structure and consisted of many of the same words. In all experiments, concrete and abstract sentences were remembered equally well when they were presented in paragraph order. When presented in random order, the usual 2:1 recall advantage for concrete materials was observed. The results are considered in terms of dual coding theory, text comprehension, and the distinction between itemspecific and relational processing. A context availability model is shown to account for the present results and a variety of other findings previously taken as evidence for and against dual coding theory.

Dimensions of metaphor

Two rating studies examined several dimensions of metaphorical sentences. A pool of 260 metaphors was constructed, all in the form “(noun phrase) is/are (noun phrase).” In Study 1 all of the items, and in Study 2, 98 of the items were evaluated on ten scales presumed to be important to the comprehension or interpretation of metaphors: semantic relatedness of the subject and predicate, comprehensibility, imageability, imageability of the subject (topic), imageability of the predicate (vehicle), degree of metaphoricity, metaphor goodness, ease of interpretation, number of alternative interpretations, and felt familiarity of the metaphoric ground. Both experiments revealed the rated dimensions to be highly interrelated, but some analyses allowed evaluation of alternative predictions based on current theoretical approaches to metaphor quality and interpretation. The results indicated consistent but mixed support for the general poisitions under consideration as each appeared to have strong and weak areas of applicability. The interrelationships among the scales are discussed, together with implications of the findings for current theories and future metaphor research.

Poetic comparisons: Psychological dimensions of metaphoric processing

A sample of 204 poetic metaphors was rated along 10 scales by 300 participants (30 different people for each of the scales). The scales were identical to ones previously used in a study involving ratings of artificially constructed metaphors, and were chosen on the basis of their relevance to current models of metaphor processing. Three major findings emerged. First, the overall pattern of findings was identical to the one obtained earlier using constructed metaphors, and aspects of it provided support for each major metaphor model without completely onfirming any one of them. Models that attribute an important role to perceptual like processes provided the most successful fit to the data. Second, all of the 10 scales were positively intercorrelated, although the correlations were generally moderate enough to permit independent experimental manipulations of the variables defined by the different scales. These results, too, are similar to those obtained earlier with constructed metaphors. Third, we identify and discuss some suggestive differences between the results of the two studies.

On memory for metaphor

Three experiments investigated how metaphors are represented in memory and the effects on memory of variables known to affect metaphor comprehension. Ten theoretically relevant dimensions were examined. In Experiments 1 and 2, free recall across a variety of orienting tasks was consistently and positively predicted by the rated imageability of metaphoric topics (sentence subjects) and the rated number of interpretations for each metaphor. The number of interpretations effect was reversed in the cued recall task of Experiment 3, as metaphors with fewer interpretations were remembered better. These and other results supported previous suggestions that memory for metaphor involves wholistic representations akin to metaphoric grounds. They also revealed consistent differences in the roles of several variables in comprehension and memory for metaphor.

Why does imagery improve memory

Abstract Evidence over the last 15 years has suggested that dual (imagery and verbal) coding explanations of concreteness effects in memory for word lists do not generalise well to memory for sentences and paragraphs. In contrast, an alternative framework based on relative differences in relational and distinctive processing has been shown to account for the effects of imagery and concreteness in these contexts and others. This paper describes recent research on free and cued recall of word lists and evaluates it with respect to the two models. The evidence suggests that whereas dual processing systems may be involved in the encoding of verbal materials, dual memory codes are insufficient to explain concreteness effects in recall. Better memory for high-as compared to low-imagery words depends on the use of paradigms that facilitate inter-item relational processing, independent of whether or not imagery is involved.

Linguistic flexibility in signed and written language productions of deaf children

Linguistic flexibility of deaf and hearing children was compared by examining the relative frequencies of their nonliteral constructions in stories written and signed (by the deaf) or written and spoken (by the hearing). Seven types of nonliteral constructions were considered: novel figurative language, frozen figurative language, gestures, pantomime, linguistic modifications, linguistic inventions, and lexical substitutions. Among the hearing 8- to 15-year-olds, oral and written stories contained comparable numbers of nonliteral constructions. Among their age-matched deaf peers, however, nonliteral constructions were significantly stories contained comparable numbers of nonliteral constructions. Among their age-matched deaf peers, however, nonliteral constructions were significantly more common in signed than written stories. Overall, hearing students used more nonliteral constructions in their written stories than did their deaf peers (who used very few), whereas deaf students used more nonliteral constructions in their signed stories than their hearing peers did in their spoken stories. The results suggest that deaf children are linguistically and cognitively more competent than is generally assumed on the basis of evaluations in English. Although inferior to hearing age-mates in written expression, they are comparable to, and in some ways better than those peers when evaluated using their primary mode of communication.

Concreteness effects in free recall: The roles of imaginai and relational processing

Two experiments reevaluated the possible role of mental imagery in free recall of concrete and abstract words. In Experiment 1, the number and rate of list presentations were manipulated. Incidental recall following an imagery rating task yielded reliable concreteness effects after two presentations but not after a single presentation, regardless of presentation rate. In Experiment 2, we examined the effects of relational (categorization) and item-specific (imagery rating) processing tasks on memory for categorically related or unrelated concrete and abstract words. Concreteness effects were obtained when unrelated words were sorted into categories but not when they were rated on imagery. Related words failed to yield concreteness effects under any orienting condition. The results support the view that the presence or absence of concreteness effects in free recall depends on the relative salience of distinctive and relational information. This conclusion constrains theoretical explanations of the role of mental imagery in memory and cognition.

Yet Another Picture of Imagery: The Roles of Shared and Distinctive Information in Memory

Consider memory for a round of golf. The event itself consists of 18 holes, but each hole is also an event consisting of different shots; on the other hand, each shot also is an event comprised by different demands. Rarely does a golfer intend to remember the events of a particular round. If, however, golfers are asked about a round, most can give appropriate accounts understandable to whomever is asking the question. For example, if asked by a colleague who has no interest in the game, one simply replies that he played golf. If asked by someone who is interested in the game, a more detailed account can be given. If asked by a suitably fanatical golfer, most golfers can give a stroke by stroke account of the previous experience. Exactly what does a golfer remember?

Semantic congruity in symbolic comparisons: Salience, expectancy, and associative priming

An expectancy interpretation of semantic congruity effects suggests that in symbolic comparisons involving the typical comparative-then-stimuli paradigm, the comparative acts as a cue in priming memory for related stimuli. A recent study by Holyoak and Mah (1981) presented evidence purported to disconfirm this hypothesis insofar as a congruity effect also was obtained when the stimuli preceded the comparative. The present study showed the stimulicomparative effect to be a consequence of the pairing of stimuli with particularly salient comparatives such that the former could serve the cuing function. This finding is consistent with the expectancy hypothesis in emphasizing the role of a flexible encoding process but inconsistent with the view that only comparatives can create expectancies in symbolic comparisons.